Thanks Manish..

Thanks Dear...

ok...where is the data ? you mentioned the name 'clientes_2.csv' but from where to get this data...

@@learndatasciencewithpranja7060 can i send you the professor files?

Thanks for Liking the video For deciding the polarity you can use either text Blob or Vader pretrained library. My experience says vader gives better results , but if your data is very specific and you think that it contains some words which are not part of English language and those words are significant in deciding if the sentiment is positive or negative then you need to train your own model

whether you take square, power 4 or 6 or 8 or 10 or any even power there shape would be same ..but while training the model you take 'Gradient' that is your slope which i hope you understand is dy/dx. when you take slope of square it would be linear if you take slope of power 4 it will be cube and cube(odd power) will have local and global minima. Similarly if original function is power 6 then when you take slope(dy/ dx) it would be power 5( odd power) and again you will have problem of local and global minima I hope you understand now....

In which institute are you studying now?

i will not suggest to use log transformation on any data where the values are zero or less than zero...as log for zero is infinity and log is not defined for negative numbers. in image data many pencells may be zeros so don't use the the log transformation

Thanks Vivek

thanks Aditya for liking the video

Hi Faris , you share your email i will contact you on your email

Pooja , My code is in python..but your code which you have used in R also seems to be correct. I can comment why its is taking so much time. Keep few points in mins while running boruta 1> No null values should be there 2> no non numeric data 3 > NO unique identifier like ( ID , sequence number etc) if you are running boruta with unique identifier like "Loan Id" column present in CR data set which i have taken in the example in this video then be assured that it will take huge amount of time please check for above three points and your issue should be resolved for sure just for your practice try running your below code on the sonar data just for your practice Sonar data is coming from the mlbench pkg. install.packages("Boruta") library(Boruta) install.packages("mlbench") library(mlbench) data("Sonar") mdl_bor <- Boruta(Class ~ ., data = Sonar) mdl_bor mdl_bor$finalDecision attStats(mdl_bor) I hope this will help

@@learndatasciencewithpranja7060 Thank you very much for ur instant reply... i tried the code u have mentioned, boruta <- Boruta(Class~., data=Sonar) , on running the above line it throwing me this error -- Error in rangerCpp(treetype, x, y.mat, independent.variable.names, mtry, : function 'Rcpp_precious_remove' not provided by package 'Rcpp' please help me with this.. for that I have installed Rcpp package also, but after the code is running...but results are not getting displayed... it simply running and taking too much of time... please help with this..please

@@learndatasciencewithpranja7060 also, on my original dataset , there are no NA's, and other ID entries, it only consist of Numeric values..but I have normalies that numeric value column...still boruta is working very slow...please help with this please....

Pooja , you have not shared the info about your data set( how many rows and cols) so its difficult for me to comment why exactly its taking so much time. While running boruta you need to keep few things in mind 1> Its a iterative algo which runs random forest for number of times so its known fact that its time consuming specially if your data has sevral thousand records and more than 30- 40 cols to reduce the time it is taking you can do sevral things check your system RAM i hope its more than 8 gb no unnecessary process are running. best thing you can do is insdead of running your Boruta on entire data run it on small % of data( first do the sampling random) which means even if run your boruta on lets say just 20 to 30 % data also property and nature of data would be same. Reducing the number of records can significantly reduce the run time you can also reduce the 'maxRuns' parameter..i guess default would be 100 you can reduce it ..but the best way would be to do random sampling and run Boruta on a small % of data instead of full data Hope this helps

@@learndatasciencewithpranja7060 my data consist of 796 obs and 25 variables..so is that a good number??to work with

Thanks for such an encouraging feedback . I have already created many videos on Algos..You can check the playlist.. will try to keep posting more videos

Thanks for liking and giving such a encouraging feedback

Thanks Jagdish...

Thanks Jagdish

Thanks Pragya